Five numbers on a screen mean nothing without a method. This is the four-step flow used by the 4D engine — and it works equally well without any software at all.
Step 1 · Observe
Record the raw numbers at warm idle. Note anything outside the normal band:
- CO > 0.5% at idle
- CO₂ < 12% at idle
- HC > 200 ppm
- O₂ > 1% with CO not-zero (mixed signal)
- NOₓ > 100 ppm at idle
Step 2 · Hypothesise (signature matching)
Match the pattern to a known signature. A handful of canonical patterns covers most real-world faults:
- Vacuum leak: high O₂, low CO₂, λ > 1.3
- Rich idle: high CO, low O₂, λ < 0.9
- Misfire: very high HC, moderate CO, elevated O₂
- Cat failure: CO/HC pass-through between pre- and post-cat (where accessible)
- EGR stuck open: rough idle, low NOₓ even loaded, moderate HC
- SAI active: high O₂ and low CO during cold phase, normalises when warm
Step 3 · Cross-check
Two independent sources that should agree:
- λ_calc vs λ_meas: Bretschneider lambda vs wideband.
- Fuel trims: STFT and LTFT should match the gas direction.
- Freeze frame: was the fault at cold start, load, deceleration?
If any of these disagree with the hypothesis, revisit before committing to a repair.
Step 4 · Confirm
Perform a physical test that directly proves the hypothesis:
- Smoke / propane test for vacuum leaks
- Holy Grail graph (lambda vs rpm sweep)
- Cylinder balance or relative compression
- Pre-post cat temperature or O₂ activity ratio
// principle
Never commission a repair on signature alone. One gas pattern can fit several faults — it's the cross-check and confirmation that makes it a diagnosis.